A robustification of independent component analysis

被引：43

作者：

Brys, G

Hubert, M

Rousseeuw, PJ

机构：

[1] Katholieke Univ Leuven, UCS, Dept Math, B-3001 Louvain, Belgium

[2] Univ Antwerp, Fac Appl Econ, B-2000 Antwerp, Belgium

[3] Univ Antwerp, Dept Math & Comp Sci, B-2020 Antwerp, Belgium

来源：

JOURNAL OF CHEMOMETRICS | 2005年 / 19卷 / 5-7期

关键词：

ICA; FASTICA; adjusted boxplot; adjusted outlyingness; skewness;

D O I：

10.1002/cem.940

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Independent component analysis (ICA) is a statistical method for transforming multivariate data to components that are as independent of each other as possible. In recent years, several algorithms were proposed that perform well in many situations. But when the data contain outliers, these methods may lead to wrong conclusions. Here we robustify the well-known FASTICA method by adding an outlier rejection rule, which does not assume elliptical symmetry. This modification is supported by simulations and real-data examples. Copyright (C) 2006 John Wiley & Sons, Ltd.

引用

页码：364 / 375

页数：12

共 26 条

[1]

[Anonymous], 2002, Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications

[2]

[Anonymous], 1987, ROBUST REGRESSION OU

[3] Normal inverse Gaussian distributions and stochastic volatility modelling [J].